Allmendinger vs. Ragan: Moneyball Turning Left

Before Best Buy took their electrons (and dollars) to RoushFenway last week, everyone assumed David Ragan was a lock for the then-empty Penske 22 seat.  Once A.J. Allmendinger landed the 22 job, there’s been a lot of speculation on the relative merits of Ragan vs. Allmendinger.  Thanks to, I was able to check some statistics.

I looked at 2011’s accomplishments using NASCAR’s Loop Data, some of which is reproduced in Table 1 below.  For comparison, I also included Tony Stewart and Carl Edwards — to get an idea of what a champion’s stats look like.

Average Start Position
Average Finish Position
Pass Differential % Quality Passes % Top 15 Laps DRIVER RATING
Allmendinger 14.9 ± 8.6 16.1 ± 8.2 3.7 ± 12.3 36.80 46.36 80.5 ± 17.3
Ragan 15.1 ± 9.5 19.2 ± 10.8 0.2 ±16.0 38.98 41.98 78.3 ± 20.5
Edwards 9.4 ± 7.5 9.3 ± 9.9 1.0 ± 12.1 64.30 70.40 101.0 ± 23.3
Stewart 17.7 ± 9.5 12.0 ± 9.9 6.6 ± 18.6 52.44 63.08 96.0 ± 25.1

Explaining the Stats

I averaged the numbers for the whole year to get an average and a standard deviation (the number after the plus or minus (±) sign).  The standard deviation tells you how much the numbers throughout the year vary relative to the average.  If someone finished in the same position every race the standard deviation would be zero.  If they finished first sometimes and 43rd sometimes and everywhere in between, the standard deviation would be large.  A small standard deviation suggestions consistency.

Pass differential is the difference between the number of times you passed someone and the number of times you were passed.  If it’s near zero, you were passed about as much as you passed other cars.  If it’s positive, you passed more cars than passed you.

Quality passes are passes of cars running in the top 15 under green.  % quality passes is how many quality passes you made divided by the total number of passes.  A high number means that you passed a lot of a top 15-cars.

% Top 15 Laps is the fraction of laps during which you are in the top 15.

Driver rating is a NASCAR mix that combines a bunch of things: whether you win, where you finish (bonus if in the top-15, bonus if you finish on the lead lap), average running position when on the lead lap, average speed under green, how often you had the fastest lap and whether you led the most laps.  The maximum driver rating per race is 150 points.  There are some older detailed explanations of this stat using the old scoring system, but I haven’t found seen any that explain it in terms of the new scoring system.  I suspect the gist is the same, with a slight modification for the 1-to-43 scoring system.

The Comparison

Two things jump out from Table 1:

1)  There’s a pretty big gap between these guys and Stewart/Edwards.

2)  Allmendinger and Ragan had pretty similar 2011 seasons – and I left out a number of other stats in which they were virtually identical.

Allmendinger has a higher average finish by three positions and spends slightly more time in the top-15.  Ragan won a race.  The one stat that is very different is that, on average, Allmendinger passes 3.7 cars more cars than pass him.  (This number represents an average over 36 races.  The number is an integer for each race because you either pass a car or you don’t, but you get a decimal when you average over the season.)  Ragan is just about even, passing as many cars as he is passed by.

Is that a big deal?  Compare Edwards and Stewart – big difference there as well.  I suspect that this has to do with the difference between your starting and finishing positions – if you start further back and finish better, you have to pass more top-15 cars.  Stewart started further back on average and had to pass more cars to get to the front.

Plate Races

I consider plate races anomalies because the basic nature is so different.  There is lot of passing, but it is very different than passing in a non-plate race.  You also have some drivers hanging in the back most of the race and then coming to the front at the end.  The average number of green-flag passes per race is about 90-100.  The average number of passes over just plate races is between 220-300. I wondered whether plate races were statistical anomalies as well, so I separated the data as shown below.

PLATE RACES ONLY Finish Pass Differential % Quality Passes % Top 15 Laps DRIVER RATING
AJ 15.8 ± 8.8 0.25 ± 13.14 41.55 32.87 70.33 ± 13.1
Ragan 20.5 ± 14.3 7 ± 32.52 67.30 57.03 90.30 ± 10.0
Edwards 14.0 ± 13.7 -10 ± 21.4 50.25 22.58 69.70 ± 16.6
Stewart 12.0 ± 3.6 2 ± 20.46 50.78 46.35 82.25 ± 21.8

Ragan’s finishes were 14, 39, 1, 28.  AJ’s were 11, 11, 10, 31.  See how consistency is reflected in the standard deviation? Just FYI:  Edwards finished 2, 6, 37, 11 while Stewart was 13, 17, 11, 7.

When you just look at plate races, Ragan’s driver rating goes up quite a bit and Allmendinger’s goes down.  But how do the stats look when the plate races are excluded?

Average Finish Position
Pass Differential Pct. Quality Passes Pct. Top 15 Laps DRIVER RATING
Allmendinger 14.8 ± 8.9 16.1 ± 8.1 4.1 ± 12.1 36.21 48.04 81.8 ± 17.4
Ragan 15.3 ± 9.1 18.7 ± 10.4 -0.6 ± 12.3 35.47 40.10 76.8 ± 20.9
Edwards 8.5 ± 7.3 8.7 ±9.2 2.4 ± 9.6 66.05 76.38 104.9 ± 21.0
Stewart 17.2 ± 9.7 12.0 ± 10.4 7.1 ± 18.3 52.65 65.17 97.7 ± 25.0

Considering that there are only four plate races, I’d give the edge to Allmendinger.

Career Trends

2011, however, isn’t the whole story, so I looked at trends over the last five years, remembering:

a)  AJ was with Red Bull his first two years, running 17/36 races in 2007 and 27/36 races in 2008.  Red Bull was a new team without top-35 owner’s points and didn’t quality for every race.  He moved to RPM in 2009 and ran all 36 races from 2009-2011.

b)  David has been with Roush/RoushFenway and ran all 36 races all fives years.

The top bar chart at right shows the number of top-five finishes and the bottom chart the number of top-ten finishes since 2007 for both drivers.  Ragan is in red and Allmendinger in blue.  The top-five bar chart really only tells us that neither of them finished in the top five consistently enough for the numbers to be meaningful.  (Edwards’ numbers for top fives for the same years are: 11, 19, 7, 9, 19.  Putting his numbers on the charge made it too hard to see the bars for Allmendinger and Ragan because they are smaller.)

The top-ten chart, however, shows Allmendinger’s consistent improvement since entering the sport.  Ragan has been more up-and-down –  a very positive blip in 2008 with 14 top tens, but he didn’t get back to his rookie year numbers until 2010. Since 2009, both are improving, but Allmendinger consistently brings in more top ten finishes.

Perhaps more telling is the second set of charts, which again show Ragan in red and Allmendinger in blue.  The top bar chart shows average starting  positions and average finishing positions are shown on the bottom chart.  Unlike the previous set of charts, lower is better .

Ragan’s standout year in 2008 again shows, but there is no question that Allmendinger’s improvement is more consistent than Ragan’s.  If you factor in that Allmendinger has been running with lower-ranked teams than Ragan, the improvement differential is even a little more impressive.

Your starting position really isn’t that important, but the graph shows a pattern of improvement, so I left it in there.

The Rest of the Field

Just for completeness, I compared the 2011 loop data states with those of  two other former Cup drivers now looking for rides.  (These are including plate tracks).  Having heard the rumors about Almirola being in line for the 43, I attempted to run his stats, but he’s run so sporadically in different series, it’s hard to make any valid comparisons.  If you compare his Nationwide stats to some of the other Nationwide drivers without rides, he doesn’t seem to me to be a good bet (which Jeff Gluck knew without having to do any math. )

Start Finish Pass Diff. Pct. Quality Passes Pct. Top 15 Laps DRIVER RATING
AJ 14.86 16.06 3.67 36.80 46.36 80.50
Ragan 15.08 19.19 0.22 38.98 41.98 78.31
Reutimann 20.44 20.58 -7.97 31.69 28.11 68.23
Vickers 19.83 20.58 -0.92 27.23 29.29 70.59

The Caveat

There’s no way to put a number on everything.  For example, RoushFenway is one of the top-tier teams, so  you would expect Ragan to have better equipment than Allmendinger.  If Allmendinger showed comparable performance in lesser-quality equipment, that might indicate potential for moving up to the next level with better equipment.  Excluding things like sponsor preferences, the stats suggest that Penske made a good decision in putting Allmendinger in the 22 – but the numbers are really pretty close.  Ragan is definitely the most proven driver out there without a Cup ride at present.




  1. Great Analysis. I appreciate a blog that’s smart, proven and doesn’t talk down to the genuine NASCAR fan. Thank you. I subscribed through RSS today. Want to read more!

  2. Great post. You did a great job explaining the statistics you posted and what they are showing.

    Looking at Reutimann’s pass differential, Clint Bowyer might be in for a long season if that’s mainly equipment based.

  3. One thing you didn’t really factor which is a big deal to car owners is busted equipment. David Ragan destroys alot of race cars…AJ is one of the least wreck prone drivers in NASCAR. Ragan’s low finished were often a result of a wreck. AJ’s bad finishes were generally the result of bad handling cars or bad luck(Dover, Indy, Texas)

  4. Cool analysis – though I’d suggest one simple improvement. As driver rating is only a calcualtion of any particular driver’s position and finish throughout a race, it would be neat to extrapolate the Driver Rating numbers a bit further to include for a “Team Strength” factor. Using one team – perhaps RFR – as a baseline, and calculating the average finishing position of all cars in any particular stable against RFR, a simple ratio for a Team Strength Factor could be easily calculated.

    As both Edwards and Ragan drive for RFR, their factor would be simply 1. But as AJ drives for RPM and as RPM’s average finish would be lower than an average RFR finish, AJ’s factor would be positive. Adding this dimension to the calcs would illustrate pretty quickly why Penske opted for AJ over Ragan.

  5. This is very good! But it also shows with Edwards and Stewart : When your hot, your HOT!!

    Great job. I would love to see more of these comparisons like maybe Jr. and Clint.



  6. good stats, worth looking at if you’re going to hire someone. What I would have been looking at is potential in Penske equipment. Ragan had a long run at Roush and only one season came close to doing anything. Allmendinger has run lesser equipment and gradually gotten better. Ragan won a race. A restrictor-plate race. It counts, but it doesn’t. Penske made the right choice. Maybe the fairest way to see things will be to judge how the new driver in the 43 does compared to Allmendinger.

1 Trackback / Pingback

  1. More Ragan vs. Allmendinger Statistics: Running at Finish »

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.